In the Hybrid Memory research approach, the development of new storing data strategies combines the traditional volatile memories (DRAM) with the emerging non-volatile memory (NVM) to take advantage of their potential and best features. One question in the hybrid memory research is the decision of which memory technology to use in order to store data of distinct applications. These decision processes have to examine the behavior of memory operations, such as read/write frequency, and also their memory characteristics. Also, considering the large volume of parameters and the uncertainties inherent in the data migration, the decision to store data in hybrid memories is not a trivial task. This paper presents the f-HybridMem component, a fuzzy-based system to support the uncertainty in data management for hybrid memory architectures. Thus, this proposal contributes to determine a correct selection between memory modules by improving the data management, in a page level organization. The memory management takes into account the behavior of memory operations and the memory characteristics. Such architecture is conceived based on the following modules: (i) Access Updater, a hardware module to identify the pages access patterns, and (ii) f-HybridMem component, a fuzzy-based software module supporting the migration decision processes. In addition, tests are conducted attempt evaluating the accuracy of the fuzzy-based system. Moreover, the proposed evaluations aim to estimate the influence of the following parameters: Buffer Size, Counters Size, Frequency of Migration, Promote Value, Demote Value, providing a correct recommend for data migrations.